Understanding congestion in airport surface operations using 3D fundamental diagrams

Abstract

Operational delays that arise when demand on the airport surface approaches or exceeds its capacity adversely impact passengers, airports, and the environment. To design effective interventions to manage airport surface congestion, airport operators require a robust understanding of the technology that drives congestion on the airport surface, that is, how delays on the airport surface vary over capacity utilization of its bottlenecks. Theoretical models of congestion technology (CT) exist, however, they are defined for ideal conditions, for instance, by assuming demand being independent of airport surface congestion, thus failing to characterize the realized airport surface operations. The availability of highly granular data on day-today surface operations facilitates the development of practically relevant models of congestion that are reproducible under wideranging operational scenarios. Nevertheless, obtaining empirical estimates of the CT from observed data on airport operations is challenging due to statistical biases that emerge via the complex interactions between air traffic operations and control at airports and in the wider airspace. In this study, we propose a novel causal statistical approach to model airport surface congestion, represented via delay versus runway and ground capacity utilization relationships, henceforth Three-Dimensional Fundamental Diagrams (3D-FDs). The proposed approach allows us to capture inherent non-linearities in the relationship while addressing the aforementioned confounding biases. Accordingly, we model the 3D-FDs of five major global airports and deliver key new insights into their surface-use efficiency, for instance, by locating their optimum operating point, that is, the point beyond which delays increase at an increasing rate with the intensity of use.

Keywords: Airports, Congestion, Congestion technology, Fundamental diagram, Causal statistical modelling, Non-parametric instrumental variables